Stanford CS229 (Autumn 2018) Problem Set & Solutions

  • Problem Set 1
    • Linear Classifiers (Logistic Regression and GDA)
    • Incomplete, Positive-only Labels
    • Poisson Regression
    • Convexity of Generalized Linear Models
    • Locally Weighted Linear Regression
  • Problem Set 2
    • Logistic Regression: Training Stability
    • Model Calibration
    • Bayesian Interpretation of Regularization
    • Constructing Kernels
    • Kernelizing the Perceptron
    • Spam Classification
  • Problem Set 3
    • A Simple Neural Network
    • KL Divergence and Maximum Likelihood
    • KL Divergence, Fisher Information, and the Natural Gradient
    • Semi-supervised EM
    • K-means for Compression
  • Problem Set 4
    • Neural Networks: MNIST Image Classification
    • Off Policy Evaluation and Causal Inference
    • PCA
    • Independent Components Analysis
    • Markov Decision Processes